Graduate Exam Abstract
May 15, 2014, 4:00 pm
ECE Conference Room (C101B)
Multifrequency Networked Microphysical Retrievals.
Abstract: This work proposes an algorithm for microphysical retrievals in a networked radar environment operating at multiple frequencies and polarizations. Microphysical retrievals are very important for both characterization of atmospheric processes, as well as ingesting into models for numerical weather prediction. Microphysical retrievals have traditionally been very diffcult however to accurately estimate and so not fully utilized. This work presents a variational algorithm for each individual instrument that can correct for attenuation errors as well as giving a best estimate of the microphysical drop size distribution. Then a bayesian multi-sensor fusion framework is demonstrated that can incorporatethe multiple radars, at multiple frequencies and polarizations, as well as non radar sources of data. This multi-sensor fusion provides a more accurate retrieval for the DSD parameters than the empirical formulas that are currently popular. This paves the way for more accurate modeling of atmospheric phenomenon.
Adviser: Venkatachalam Chandrasekar
Non-ECE Member: Paul Mielke, Statistics
Member 3: Anura Jayasumana, ECE
Addional Members: Margaret Cheney, ECE/Math
Program of Study: